An extended Kalman filter for in situ sensing of yttria-stabilized zirconia in chemical vapor deposition

Chemical vapor deposition (CVD) is an important step in integrated circuit fabrication and is a dynamic process that is complex and subject to unmeasured disturbances. Currently in semiconductor industry the main form of advanced process control for CVD is run-to-run (R2R) control. R2R control is a form of discrete process control in which the processing recipe is modified ex situ based on post-process metrology data. Ideally one would like to monitor and control film growth in situ during deposition. To achieve this goal an in situ sensor that can sense film properties in real time is needed. This paper studied an extended Kalman filter (EKF) based in situ sensor and demonstrated its application in chemical vapor deposition of yttria stabilized zirconia (YSZ). The sensor consists of a reflectometer, a data preprocessing module, and an EKF algorithm. The reflectometer measures reflected light intensity from the film surface at two wavelengths, 950 nm and 470 nm. The data preprocessing module calibrates raw intensity data, estimates the initial state and triggers EKF. EKF estimates film thickness, roughness, and growth rate by minimizing the mismatch between the estimated and measured reflectance. The estimated film properties were compared with ex situ characterization using scanning electron microscopy (SEM) and atomic force microscopy (AFM). Sensitivity and robustness of EKF were also studied by both simulation and experimental implementation.